Introduction
The field campaigns of the European Commission Mesometeorological Cycles of Air Pollution in the Iberian Peninsula (MECAPIP) project and, in particular, the pollutant distribution and meteorological data associated with the July 1989 flights of an instrumented aircraft produced important results that have changed the state of knowledge of the spatial distribution and dynamics of pollutants in the Iberian Peninsula radically. These results have been published in various journals and other specialized media (Martin et al. 1991; Millán et al. 1991, 1992, 1996, 1997).
In addition to the flights over the Spanish Mediterranean coast and the flights from the east coast to Madrid (Millán et al. 1992), the measurement campaign also included a series of flights over the northern coast of the Iberian Peninsula on 25 and 28 July (Fig. 1 shows the Iberian Peninsula with its main topographic features). These flights followed a north–south route from Bilbao to Madrid (on 25 July) and a southeast–northwest route over the Ebro valley (on 28 July). Both of these flights included a closed loop along the northern coast. Vertical layering of the air pollutants observed during these flights showed specific features with respect to the rest of the data obtained over the Iberian Peninsula: the stratification remained unchanged during the morning and afternoon flights (Millán et al. 1992), and the persistent northerly winds of the upper layers, placed above the sea–land breeze, indicated a different origin for those polluted air masses.
Thus, the aim of this paper is to describe the complex atmospheric circulation of the lower troposphere often found in summer over the northern coast of the Iberian Peninsula and to search into the mechanisms leading to the kind of layering observed in the polluted air masses. A detailed analysis of the flights over that coastal area has been included as well as back trajectories computed by using a 3D wind field simulated by the Regional Atmospheric Modeling System (RAMS) to search for sources of pollutants. The data from 25 July, considered to be representative of a typical summer scenario, were selected for processing in this paper.
Experimental results
The pressure pattern over the Iberian Peninsula on 25 July can be considered to be typical of the summer period, and Fig. 2 shows the map of the geopotential function for the 1000-hPa surface, corresponding to the 1200 UTC analysis from the European Centre for Medium-Range Weather Forecasts (ECMWF). The Azores anticyclone extends a ridge of high pressure over the Cantabrian coast that covers the entire Bay of Biscay, and over the Iberian Peninsula a cycle of thermal lows that develop during the day and relax at night is established (Millán et al. 1997).
To obtain successive vertical transects through the polluted air mass, the flight path was of the “sawtooth” (or dolphin) type and included consecutive ascents and descents from approximately 50 m above ground to the flight ceiling. This procedure, with the kind of aircraft used (a twin-jet Hawker Siddeley HS 125), allowed for flight times on the order of 1 h or less from the center of Spain to any of the coasts, with as many as 11 legs (profiles) in the 350–400-km distance. The Fraunhofer Institute for Atmospheric Environmental Research (Garmisch–Partenkirchen) and the Aerodata Co. from Germany operated the instrumented aircraft; flight routes and a complete description of the intensive field campaign, equipment, and data handling can be found in Millán et al. (1992, 1996). The ground track for the 25 July flight (1114–1246 UTC) and the main orographic features are shown in Fig. 3.
The flight crossed the upper Iberian plateau from Madrid at 1114 UTC, near the lower point “L” of the trajectory in Fig. 3, to the Bilbao airport, marked “S,” at 1246 UTC. The average terrain heights at the plateau are between 600 and 700 m above sea level (MSL). After descending to the northern coast, it completed the “O–P–W” loop before landing at the airport. The city of Bilbao is located in a basin estuary that still has important industrial activity, including foundries, refineries, paper mills, fertilizer and other chemical plants, and a power plant that operates intermittently during peak periods of electricity demand.
Figures 4 and 5 show the distribution of ozone (O3) and reactive nitrogen (NOy), respectively, together with the flight legs and terrain profile. On the vertical scale the units represent meters above sea level, and the horizontal scale shows kilometers from point L in Fig. 3. The gray-scale code and contour lines indicate the concentrations of the different gases in ppb(v) interpolated between the ascending and descending trajectories of the aircraft. All values are generated around the flight legs, following a severe distance-dependent interpolation scheme. The distribution of ozone above 1500 m and the simultaneous concentration levels of NOy and NOz [NOy − (nitric oxide + nitrogen dioxide)], which are depicted in Fig. 8, correspond to a relatively aged, polluted air mass located over the eastern part of the route, covering the Bay of Biscay and the vertical region over the estuary of Bilbao. The levels of NOy for the lower layers and those of sulfur dioxide (SO2) in Fig. 6 coincide in the same area and indicate the presence of new emissions trapped at the bottom of the valleys of the coastal area. This volume is part of a blocked air mass formed when the drainage flows reach the sea.
Figure 7 shows vertical profiles of meteorological data obtained by the aircraft along the flight leg marked “A–B” in Figs. 4–6, which corresponds to the descent from the plateau to the coast; wind, dewpoint, and dry and potential temperatures are illustrated. The concentrations of different gases for the same vertical section are shown in Fig. 8, as is the photochemical age (NOz/NOy). Descent of the aircraft from point A, in Figs. 4–6, was interrupted by an ascent from 500 m MSL and then again a descent to sea level; this pattern caused “false double values” of data between 500 and 800 m MSL in Figs. 7 and 8. The relationship between the vertical stratification of the atmosphere and the distribution of pollutants can be observed clearly. The mass of aged pollutants is associated with north-northeast winds between an upper-level temperature inversion and a near-isothermal layer located at about 1200 and 2000 m, respectively. There is no nitric oxide (NO) in the upper layer, and the concentrations of SO2 and NOy are low and have the same spatial distribution as does ozone, which peaks at higher concentrations—almost 80 ppb—just above the upper inversion. The corresponding ratios of NOz:NOy are at their highest values.
There are two layers below the temperature inversion at 1200 m shown in Fig. 8. The one closest to the surface, below 300 m at the coastline, includes an intense sea breeze (over 5 m s−1 in Fig. 7) with a north-northwest component that creates its own temperature inversion as it enters the valleys, displacing and replacing the air mass that has come from nocturnal drainage flows and blocking. This blocked air mass, lifted up by the sea breeze, forms the second (lower) layer, and at this time (∼1200 UTC) it remains trapped below the subsidence inversion at 1200 m. It still contains recent industrial emissions of local origin, which include clearly defined peaks of SO2, NO, and nitrogen dioxide (NO2), and the ratio NOz:NOy displays the lowest values.
The afternoon flight (1445–1615 UTC), shown in Fig. 9, had an almost identical trajectory but ran in the opposite direction, from Bilbao (S) to Madrid (L), after completing the “S–P–O” loop over the northern coast. The concentrations of O3 and NOy (Figs. 10 and 11) show the same vertical layering or stratification as they did in the morning, and the concentrations persist in the eastern area of the route over the sea and inland up to the plateau. In this area, between 200 and 250 km from point L and close to the mountains of the Cantabrian Range that separate the coastal area from the plateau, the ozone transported in the upper layers is fumigated to the ground. By this time of day, the pollutants emitted on the coast during the night and morning of the same day also are flowing inland in the lower layers within a fully developed sea breeze. Thus, in the fumigation region, the upper and lower air masses are combined, that is, the O3 from long-range transport mixes with the pollutants emitted on the coast.
Figure 12 shows the meteorological data for the leg marked “C–D” in Figs. 10 and 11. It corresponds to the ascent from the coast to the plateau. As the profiles for the dewpoint and temperature indicate (Fig. 7), the anticyclonic subsidence inversion, which was located at 1200 m in the morning, has sunk to 900 m over the Cantabrian coast. Above that level the wind still maintains a marked northerly component, which explains the displacement of the pollutants in the upper layers. Nearer to the surface, the blocking in the valleys has been replaced by a moderate to strong breeze toward the interior.
Figure 13 illustrates the profile obtained by the instrumented aircraft over the plateau, along the leg “E–F,” shown in Figs. 10 and 11. The temperature and dewpoint indicate that the mixing layer reached a depth of at least 2200 m at this distance inland. This growth is due to strong convection over the upper Iberian plateau, which explains the fumigation and mixing of the fresh pollutants in the lower layers with those transported at the higher levels.
To sum up, the evolution observed on the surface and on the Cantabrian north-facing slopes is characterized by a cycle of drainage flows and blocking over the coast in the morning and reentry of locally emitted pollutants within the sea breeze in the afternoon, while at a higher level a layer of older pollutants and high O3 concentrations is being advected continuously by winds with a northerly component. This flow structure has been confirmed further by the meteorological records of the automatic surface network of the Basque Meteorological Service and of the Spanish Meteorological Service and the data from the soundings at 0000 and 1200 UTC of the Santander regional center, which is the closest to the Bilbao area; in this case, the persistence of the winds with a northerly component at the higher level, not only on 25 July but also on the two previous days, suggested the need to search for the source of the pollutants that appear in the upper levels of the flight routes over the northern coast and are transported inland over the peninsula.
Meteorological simulation and back trajectories
A back-trajectory technique was used for this search and was applied with wind data with a sufficient spatio-temporal resolution. The gridded hemispheric analysis from ECMWF has a resolution of 1°, which is equivalent to approximately 100 km, and it was decided to increase the resolution to 9 km through use of a mesoscale meteorological simulation model. The third version of RAMS from the University of Colorado (Pielke et al. 1992) was used to carry out the simulation. It is a nonhydrostatic prognostic model and was initialized using the ECMWF analysis. Because convection, coastal recirculation, and orographically aided injection processes can be simulated explicitly at the selected grid resolution, the three coordinates u, υ, and w of the wind vector were used to track better the trajectories of the air parcels. The back-trajectory concept does not include diffusive transport mechanisms, but this lack is of less importance when the transport occurs under stable conditions over the sea, as is the case here.
Two nested grids were used, with 9- and 27-km resolution in the horizontal dimension and variable resolution in the vertical, from 0 to 13 000 m in height, with minimum increases of 50–100 m close to the ground and maxima of 1000 m at the higher levels. Two-way nesting was allowed between the grids, so that the finer grids conditioned the flow in the coarser grids and vice versa. The four-dimensional data assimilation technique was used for the simulations, nudging the boundary and center of the grids to the conditions established by the ECMWF analysis data at 0000, 0600, 1200, and 1800 UTC. To proceed with a back-trajectory estimation of a sufficiently wide time interval, the run extended over 24 and 25 July, covering all 24 h of 24 July and the first 18 h of 25 July. The fine grid included the northern part of the Iberian Peninsula, and the coarse grid covered the whole peninsula plus the north of Africa and a large part of France, up to the Brittany peninsula.
The upper part of Fig. 14 shows the simulation of the wind field at the surface at 0600 on 25 July in the high-resolution grid. The representation is based on the topography “interpreted” by the model itself on a gray scale for which the levels (in meters) appear in the lower part. The drainage flows toward the sea are simulated well over the north-facing Cantabrian range, as are the drainage of the Ebro valley into the Mediterranean Sea and the Duero valley into the Atlantic Ocean. A northerly wind blows in the Bay of Biscay, joins the drainage of the valleys, and is deflected eastward along the coastline, and in the northwestern corner of the Iberian Peninsula the wind increases to continue parallel to the coast of Portugal.
The lower part of Fig. 14 shows the surface wind field at 1400 on the same day. The coastal breezes and the coupling with the higher-level wind of northerly component generate this more “homogeneous” wind field with clear channeling along the Ebro and Duero valleys. The evolution of the wind at the higher level does not present as many changes, as can be seen in Fig. 15, which shows the wind field at 1500 m above the ground for the same times as in Fig. 14.
The back trajectories were traced using the simulation output data on both grids. Initially, the wind data from the high-resolution grid were used, as shown in Figs. 14 and 15. When the trajectories leave the domain, the wind data from the low resolution grid were utilized (Fig. 16) to obtain a wider spatial coverage. The computation was carried out over a period of 36 h: 24 h on 24 July and 12 h on 25 July. The wind field was updated every 2 h (the recording period chosen for the meteorological simulation data).
Figure 16 shows the plan view of the back trajectories, which ended at 1200 UTC between 1400 and 1700 m over the vertical of Bilbao. The topographic levels interpreted by the simulation model on the coarse grid—with wider spatial coverage—are represented on a gray scale similar to that used in Fig. 1. Figure 17 presents a vertical view of the same trajectories as in Fig. 16, projected along a meridian section at 3°E longitude. The gray scale here represents the vertical wind speeds in m s−1 at 1200 UTC on 25 July, which corresponds to the arrival time of the back trajectories. It can be observed that the air undergoes a clear process of sinking over the Bay of Biscay (down to 700 m MSL for one of the trajectories) and ascent in the coastal areas. These processes determine the vertical movement of the air mass that comes from the north and reaches the eastern corner of the north coast of the Iberian Peninsula.
It can be concluded, therefore, that the layer of old pollutants that appeared on 25 July over Bilbao originated from emissions released in the region near the English Channel or from further away. The possibility of long-range transport of European pollutants toward the Atlantic in the northeastern branch of the Azores anticyclone also has been examined by Flatøy et al. (1995), and the data presented here support this kind of process.
Conclusions
A detailed analysis of the concentrations of O3 and other characteristic pollutants and their vertical stratification over the eastern Cantabrian coast measured during the experimental campaigns carried out within the MECAPIP project demonstrates the long-distance transport of air masses with high concentrations of ozone of continental origin over that area. These pollutants were detected under typical summer conditions, with an intense anticyclone over the Azores and a ridge of high pressure over the north coast of the Iberian Peninsula, and appear in the upper layers of the atmospheric soundings over the coast of Bilbao. They are separated from the lower layers by a well-marked (anticyclonic) subsidence inversion.
During night and early-morning conditions, the lower air mass at the vertical of the Bilbao estuary, below the anticyclonic subsidence inversion, remains decoupled from the above-mentioned higher layers. This lower air mass initially includes the drainage flows, which tend to get blocked upon reaching the sea, and later during the day includes the developing sea breeze. This lower air mass also accumulates and transports the pollutants emitted locally during the day and previous night. Later in the day, the upper and lower layers are observed to couple and mix as soon as they move inland. This coupling occurs at the same time as the appearance of the sea breezes on the coast and the development of a series of convective cells over the central plateaus that make up the mosaic of the Iberian thermal low.
Along the Spanish northern coast there is evidence that this coupling and mixing process takes place by means of the thermal convective cells that develop on the slopes of the mountains in the Cantabrian Range and over the upper plateau of the Iberian Peninsula. At about the same time, during the afternoon, a strong sinking of the subsidence inversion has been observed to occur over the coast near Bilbao, which intensifies the decoupling between the atmospheric layers above the subsidence temperature inversion from those in the lower layers involved in the sea-breeze cycles.
The convective mixing within the upper plateau thus contrasts with the increasing stabilization observed over the north coast, caused by the intensification of compensatory sinking over the colder sea. This kind of process has been further confirmed to occur frequently during the summer period, by the data from a boundary layer wind profiler operating at the inlet of the Bilbao estuary since 1996 (Alonso et al. 1998).
Acknowledgments
The results presented in the present study were obtained within the framework of the European MECAPIP project. We would like to thank the Centre for Environmental Studies of the Mediterranean for access to the RAMS results. We also are grateful to the people who participated in the field measurement campaigns, without whose collaboration it would not have been possible to obtain the extensive database of results on which this piece of work is based:José Antonio García, Nieves Durana, Juan Luis Ilardia, and Cristina Gutiérrez Cañas, among others.
REFERENCES
Alonso, L., G. Gangoiti, M. Navazo, M. Maruri, J. A. García, and J. A. Aranda, 1998: The Punta Galea boundary layer profiler: Intercomparison with radiosonde data and first mesoscale meteorological case studies. Meteor. Z.,7, 203–212.
Flatøy, F., Ø. Hov, and H. Smit, 1995: 3-D model studies of vertical exchange processes of ozone in the troposphere over Europe. J. Geophys. Res.,100, 11 465–11 481.
Martín, M., J. Plaza, M. D. Andrés, J. C. Bezares, and M. M. Millán, 1991: Comparative study of seasonal air pollutant behavior in a Mediterranean coastal site: Castellón (Spain). Atmos. Environ.,25A, 1523–1535.
Millán, M. M., B. Artíñano, L. Alonso, M. Navazo, and M. Castro, 1991: The effect of meso-scale flows on the regional and long-range atmospheric transport in the western Mediterranean area. Atmos. Environ.,25A, 949–963.
Millán, M. M., B. Artíñano, L. Alonso, M. Castro, R. Fernández-Patier, and J. Goberna, 1992: Meso-Meteorological Cycles of Air Pollution in the Iberian Peninsula (MECAPIP). Air Pollution Research Rep. 44, EUR No. 14834, 219 pp. [Available from Communication Unit, European Commission DG XII/E-1, Rue de la Loi 200, B-1040 Brussels, Belgium.].
Millán, M. M., R. Salvador, E. Mantilla, and B. Artíñano, 1996: Meteorology and photochemical air pollution in southern Europe: Experimental results from EC research projects. Atmos. Environ.,30, 1909–1924.
Millán, M. M., R. Salvador, E. Mantilla, and G. Kallos, 1997: Photooxidant dynamics in the Mediterranean basin in summer: Results from European research projects. J. Geophys. Res.,102, 8811–8823.
Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modelling system—RAMS. Meteor. Atmos. Phys.,49, 69–91.